HomeCyber AttacksGoogle’s AI brokers are already stopping real-world cyberattacks

Google’s AI brokers are already stopping real-world cyberattacks


Readers assist help Home windows Report. We might get a fee for those who purchase by our hyperlinks.

Learn our disclosure web page to search out out how will you assist Home windows Report maintain the editorial crew. Learn extra

Google is moving into this summer season’s cybersecurity convention season with a wave of updates, demos, and real-world wins for AI-powered security. In a brand new weblog publish, the corporate shared how its agentic AI instruments, just like the Massive Sleep agent, are already figuring out vulnerabilities earlier than hackers can exploit them.

Massive Sleep, first introduced final 12 months, just lately discovered a crucial SQLite flaw (CVE-2025-6965) that was beforehand identified solely to menace actors. With intel from Google’s Menace Intelligence crew, the AI predicted an imminent assault and blocked it preemptively, probably a primary in public cybersecurity historical past. That very same AI now helps open-source security, serving to plug holes throughout the broader web.

See also  DHS Warns Professional-Iranian Hackers More likely to Goal U.S. Networks After Iranian Nuclear Strikes

At Black Hat USA, Google plans to demo new AI capabilities together with Timesketch (an incident response device powered by Sec-Gemini) and FACADE, which detects insider threats throughout billions of occasions a day. The corporate may also be a part of Airbus for a DEF CON Seize the Flag occasion pairing human gamers with AI assistants.

Google is donating information from its Safe AI Framework to assist drive industry-wide security efforts beneath the Coalition for Safe AI. And subsequent month, it can wrap up its two-year AI Cyber Problem with DARPA, the place finalists will exhibit instruments constructed to safe open-source software program.

Whereas the tech itself is spectacular, Google says none of it issues with out accountable deployment. The corporate is pushing secure-by-design practices, transparency, and human oversight to make sure AI stays an asset, not a legal responsibility.


See also  4 Arrested in £440M Cyber Attack on Marks & Spencer, Co-op, and Harrods

- Advertisment -spot_img
RELATED ARTICLES

LEAVE A REPLY

Please enter your comment!
Please enter your name here

- Advertisment -

Most Popular